Distributed meta-heuristic method for network optimization problems in information network

This paper deals with a network optimization problem with the response-time limitation that minimizes the total capacities of communication channels for a broadcast communication from a given source disseminating information in a network. The solving method and the solution of this problem are defined as E-Transformation and E-Tree. But finding E-Tree that is the minimum cost tree-shaped network essentially belongs to class NP-hard, and we propose meta-heuristic approaches, i.e. Local Search and Tabu Search, with the restriction of the height of a tree for precise approximate solutions. The condition of the height of a tree, which is a property of the network topology, restricts the searching range considerably. We also propose a distributed algorithm of E-Tree search for the sake of an application as an intelligent protocol in a practical large scale information network